[11:04:50] https://t.me/bananadeployer [11:44:41] Two spammers it seems [11:49:44] ooh, apparently Telegram can now delete reactions by spammers [11:49:59] pretty sure someone was complaining about that the other day, not sure if it was here or in the Hackathon group [12:47:20] Is this a good place to get help on why two links are missing from the external links table? [12:47:26] https://quarry.wmcloud.org/query/105275 [12:47:55] More context https://en.wikipedia.org/wiki/Wikipedia_talk:Archive.today_guidance#c-WhatamIdoing-20260513074400-Missing_some_links [12:50:01] I was! (re @lucaswerkmeister: pretty sure someone was complaining about that the other day, not sure if it was here or in the Hackathon group) [12:51:05] Dw31415: the action API also only sees one extlink to archive.today https://en.wikipedia.org/w/api.php?action=query&prop=extlinks&ellimit=max&titles=Aktion%20T4&formatversion=2 [12:51:34] looking at the source code, I notice that only one archive.today link is a [plain link]; the other mentions go through {{cite web}} where it’s used as |archive-url= [12:51:55] my guess is {{cite web}} (or one of its nested templates) has been made to not link to archive.today, because of the shady stuff the operator of that archive site pulled [12:52:16] or at least I don’t see an archive link on those rendered references on the article [12:52:23] That is a good point. What’s odd is changing to eswiki_p shows all 3 links [12:52:44] different wiki, different templates? [12:52:46] different wikis have different content… [12:53:54] Maybe the external links pick up after template rendering? [12:54:22] The CS1 module is hiding the links from rendering [12:54:28] https://en.wikipedia.org/w/index.php?title=Module:Citation/CS1&diff=prev&oldid=1340931214 appears to be the diff that made the module not link to archive.today [12:54:37] (though the edit summaries at https://en.wikipedia.org/w/index.php?title=Module:Citation/CS1&action=history are infuriatingly unhelpful in general) [12:54:53] Dw31415: yes, it’s external links in the rendered output [12:55:03] not strings that happen to look like URLs in the wikitext input [12:55:10] Well that answers it! [12:55:41] Is there a way to query for the links in the source? [12:56:09] So we can have reports that include the invisible links? [12:56:26] search for insource:"archive.today", I guess? [12:57:08] What table is that? [12:57:16] no table. search [12:57:29] https://en.wikipedia.org/w/index.php?search=insource%3A%22archive.today%22&ns0=1 [12:58:07] Well you answered my question. Thanks! [12:58:13] theoretically, enwiki's CS1 templates might also be able to be configured to add a page to a tracking category when they detect an archive.today link [12:58:43] I think it does generally… [12:58:44] good point [12:59:01] The user wanted a project specific report. [12:59:43] Thanks again! [13:01:49] Dw31415: do you mean a WikiProject specific report? in that case i haven't looked deeply into it, but you might potentially be able to use something like https://petscan.wmcloud.org/ to find the intersection between articles in https://en.wikipedia.org/wiki/Category:CS1_maint:_deprecated_archival_service and articles in a WikiProject's category [14:23:06] !log lucaswerkmeister@tools-bastion-15 tools.ranker deployed 9c42483353 (l10n updates: ko) [14:23:08] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.ranker/SAL [14:24:04] !log lucaswerkmeister@tools-bastion-15 tools.wd-image-positions deployed 5e96e1a7f9 (l10n updates: ko, vi) [14:24:05] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.wd-image-positions/SAL [15:00:18] Hi all, I am a student developing a tool on Toolforge for my master thesis. I created a user database with ToolsDB and my current table holds about 25mio Wikidata entities. Does maybe anyone here have tips/tricks/resources or knowledge on how to optimize database architecture and queries for this amount of data? My first test endpoint took 10 minutes for an ok and I wonder [15:00:18] what's [15:00:18] the next best approach to reduce it.. [15:00:53] share your schema and typical queries? [15:33:13] Hi folks. Two questions, if I may i) what's the expected turnaround for quota tickets (T425975) ii) can I get debian-11.0-bullseye image available again, please? I tried re-activating one of the bullseye images via horizon, but get a permission error. Backround is I want to test the bullseye->trixie migration for production swift, which is currently still on bullseye, so I need it as a starting point. [15:33:14] T425975: Quota increase request for project swift - https://phabricator.wikimedia.org/T425975 [15:36:24] Emperor: for i) that's taken care by the clinic duty person (and coincidentally you're asking exacttly at the time of the weekly meeting after which it changes) [15:37:09] for ii) I'm not sure, leaving the answer to those that know more [15:38:49] volans: thanks. Your answer to i doesn't tell me what the SLO (for want of a better phrasing is), only who is responsible :) [15:39:05] s/ is)/) is/ [15:40:36] I don't think there is an "SLO" for them, but I see you already got a +1 so it shouldn't take long, I guess today/tomorrow [16:35:23] How are you storing the data in the database currently? (re @Uyen: Hi all, I am a student developing a tool on Toolforge for my master thesis. I created a user database with ToolsDB and my curren...) [16:36:24] Emperor: I'm probably the one who can re-enable Bullseye; do you want an image that attempts to puppetize itself, or just a raw bullseye cloud image that you manage with cloud-init? (I imagine that puppet will be mostly broken on Bullseye these days). [16:39:26] volans: is this a clinic duty within wmcs or do you mean the SRE clinic duty? [16:42:41] wmcs [16:42:58] gotcha. it was curiosity because I was just in a meeting talking about clinic duty [16:43:13] not about the specific request. thanks! [16:43:40] Historically we reviewed quota requests weekly. So in my mind the 'SLO' is always 'by next Thursday' and we pretty much always exceed it unless the request is controversial. [17:34:10] hey, just out of my own personal curiosity, i was wondering if there'd been a change to the DNS setup for *.wmcloud.org domains recently. i could be misremembering, but IIRC, previously going to a subdomain like https://doesntexist.wmcloud.org/ would show you a generic Cloud VPS error screen, but now a nonexistent *.wmcloud.org subdomain like that doesn't seem to have any DNS records at all FWICS [17:34:25] nonexistent domains on *.wmflabs.org still show a Cloud VPS error screen FWICS: https://doesntexist.wmflabs.org/ [17:45:48] A_smart_kitten: a generic star record has never been there on wmcloud.org [17:46:14] huh, maybe i am misremembering then :) [17:46:33] maybe i was getting wmcloud.org confused with wmflabs.org in my mind without realising it [17:46:39] toolforge does that, as well as wmflabs.org (to support the auto-redirector), but that domain does not [17:46:40] thanks for the response taavi! [17:47:27] ahh maybe i was also getting confused with toolforge without realising it [21:13:39] Hello BeeBee_P :) [21:50:45] Thank you and taavi for your responses. I created a snippet for this https://phabricator.wikimedia.org/P92550 (re @harej: How are you storing the data in the database currently?) [22:51:57] andrewbogott: one that puppetises, I think (this is to use in a pontoon environment - https://wikitech.wikimedia.org/wiki/Puppet/Pontoon) analogous to e.g. debian-12.0-bookworm - bullseye puppet definitely still works across all of ms-* and thanos-* in production still... [22:58:14] andrewbogott: really, I think "whatever the current debian-12.0-bookworm image does; and likely thus whatever the now-disabled debian-11.0-bullseye image did" [23:19:18] ok, let's see what I can scare up here... [23:19:33] what project? [23:37:27] Emperor: what project do you need the Bullseye image for?